Ecological Indicators 23 (2012) 66–75

Contents lists available at SciVerse ScienceDirect

Ecological Indicators
journal homepage: www.elsevier.com/locate/ecolind

Benthic foraminifera provide a promising tool for ecological quality assessment

of marine waters
Vincent M.P. Bouchet a,∗ , Elisabeth Alve a , Brage Rygg b , Richard J. Telford c
a
Department of Geosciences, University of Oslo, P.O. Box 1047, Blindern 0316 Oslo, Norway
b
Norwegian Institute for Water Research (NIVA), Gaustadalèen 21, 0349 Oslo, Norway
c
Department of Biology, University of Bergen, Thormøhlensgate 53 A og B, 5006 Bergen, Norway

a r t i c l e i n f o a b s t r a c t

Article history: This study defines criteria for the use of benthic foraminifera (protists) as a quick and efficient bio-
Received 25 July 2011 monitoring tool to implement marine legislation. Various sampling and preparation procedures are
Received in revised form 7 March 2012 investigated in an attempt to find the optimal methodology for environmental monitoring using soft-
Accepted 14 March 2012
sediment foraminifera with the objective of assessing ecological quality status (EcoQS). Twenty-seven
sampling stations in silled basins along the Norwegian Skagerrak coast, NE North Sea, are investi-
Keywords:
gated for environmental parameters and living (stained, including soft-shelled forms) and dead benthic
Environmental monitoring
foraminifera. Diversity, expressed as the effective number of species (exp H bc ) and community compo-
Ecological quality status
Living benthic foraminifera
sition are used to evaluate EcoQS using living (stained) benthic foraminifera. Correlation studies show
Oxygen gradient that bottom-water dissolved oxygen concentration at the time of sampling ([O2 ]tos ) is the main envi-
Diversity index ronmental factor controlling variation in diversity. Variables such as grain size, C/N, TOC and TN are less
Methodological procedures important. Correlation between foraminiferal diversity and [O2 ]tos , as well as correlation between com-
munity data and [O2 ]tos , suggest that benthic foraminifera represent an efficient bio-monitoring tool to
evaluate EcoQS. A clear pattern from “bad” to “high” EcoQS is established using the strong link between
the benthic foraminiferal diversity and the bottom-water oxygen gradient. The study shows that EcoQS
can be evaluated quickly and accurately using the following method: sample the top 1 cm of sediment,
dry-pick about 250 living (stained) individuals of >125 ␮m-sized fossilisable (i.e. most of those remain-
ing subsequent to drying) foraminifera from each of three replicates. exp(H bc ) based on living benthic
foraminifera is a promising tool to assess EcoQS. For fossil assemblages, exp(H bc ) has potential for eval-
uating temporal changes in in situ PaleoEcoQS and for defining reference conditions from pre-impacted
times.
© 2012 Elsevier Ltd. All rights reserved.

1. Introduction of ground, continental, transitional, and coastal waters, emphasis-

Since the end of the 19th century, human activities have greatly
altered the ecological quality of many coastal areas. In response
to concerns about environmental degradation, many nations have
enacted legislation such as, for instance, the Clean Water Act
(CWA) or Oceans Act in USA, Australia or Canada, the Water
Framework Directive (WFD, 2000/60/EC) and the Marine Strate-
gic Framework Directive (MSFD, 2008/56/EC) in Europe, to address
anthropogenic pollution. The CWA established the goals of elim-
inating releases of large amounts of toxic substances into water,
eliminating additional water pollution, and ensuring that sur-
face waters would meet standards necessary for human sports
and recreation. The WFD established a basis for the protection
∗ Corresponding author. Tel.: +47 22 85 49 08; fax: +47 22 85 42 15.
E-mail address: vincent bouchet@hotmail.fr (V.M.P. Bouchet).
ing the need to monitor and assess the ecological quality status
(EcoQS) of those ecosystems. According to the WFD, European
countries must restore the environment to a “good” EcoQS by
2015. The MSFD aims to “achieve good environmental status of
the EU’s marine waters by 2020 and to protect the resource base
upon which marine-related economic and social activities depend”
(www.ec.europa.eu/environment/water/marine.htm). The MSFD
highlights the need for the scientific community to increase the
scientific knowledge of the elements that define the state of the
marine environment. The implementation of these marine legis-
lations worldwide is generating a fruitful debate amongst marine
scientists about how to define and implement efficient and reli-
able bio-assessment tools. In coastal and transitional waters, many
biological elements have been considered as assessment tools,
including macroalgae, phytoplankton and sea-grasses (Ballesteros
et al., 2007), fish (Coates et al., 2007), and benthic macrofauna (see
review in Pinto et al., 2009). Monitoring programs in soft-bottom

1470-160X/$ – see front matter © 2012 Elsevier Ltd. All rights reserved.
doi:10.1016/j.ecolind.2012.03.011
 V.M.P. Bouchet et al. / Ecological Indicators 23 (2012) 66–75
coastal and transitional waters have highlighted the difficulty of
defining reference conditions (e.g. Bigot et al., 2008; Blanchet
et al., 2008; Elliott and Quintino, 2007; Nielsen et al., 2003).
According to the WFD, the reference conditions (also called “high”
status) are defined as “. . . the biological, chemical and morpho-
logical conditions associated with no or very low human pressure.
These reference conditions are type-specific, so they are differ-
ent for different types of rivers, lakes or coastal waters so as
to take into account the broad diversity of ecological regions in
Europe. Assessment of quality is based on the extent of deviation
from these reference conditions, following the definitions in the
Directive” (www.ec.europa.eu/environment/water/marine.htm). It
means that a low-diversity assemblage which would be classified
as reflecting “poor status” at one site may actually be the reference
condition (and should be classified as “high status”) at another site.
A recurring problem is that, unfortunately, for most areas, there
are no data available from pre-impact times to use as reference
conditions, and with which to compare the current EcoQS. Refer-
ence conditions are therefore defined using expert judgment. Even
though expert judgment may have some advantages, a more objec-
tive approach is needed.
A component of the meiofauna, benthic foraminifera (protists),
has been widely used as bio-indicators of several pollution sources
in coastal and transitional waters (e.g. Alve, 1995; Martínez-Colón
et al., 2009; Nigam et al., 2006; Scott et al., 2001) such as aquacul-
ture (e.g. Bouchet et al., 2007), oil spills (e.g. Morvan et al., 2004),
heavy metals (e.g. Armynot du Châtelet et al., 2004), and urban
sewage (e.g. Burone et al., 2006). Benthic foraminifera occur in
almost all marine environments, and with much higher abundances
than macrofauna, which is traditionally used in environmental
monitoring. Foraminifera can provide quantitative data with a
small (a few cm3 ) volume of sediment. Considering their shorter
generation time, foraminifera have the potential to respond faster
than macrofauna to changes in the environmental conditions. Stain-
forthia fusiformis, one of the most abundant species in Northern
European fjordic systems, is, for instance, able to reproduce within
less than a month in response to food input (Gustafsson and
Nordberg, 2001). In the deep sea, benthic foraminifera react within
1–2 months to the deposition of phytodetritus with dramatic
changes in the assemblage composition (Gooday, 1988). Benthic
foraminiferal assemblages have also been able to recover their
composition and abundance within a month after a short-term
hypoxic period in intertidal mudflats in S-Western France (Bouchet
et al., 2007). Benthic foraminifera leave an easily accessible and
abundant fossil record, which allows reconstruction of the char-
acteristics and timing of historical environmental variations (e.g.
Alve, 1991a; Debenay and Fernandez, 2009; Hayward et al., 2004).
Consequently, it is possible to trace the record of human-induced
disturbances over decades or centuries. Indeed, in a pilot study,
Alve et al. (2009) suggested that in situ reference conditions can
be established using fossil benthic foraminiferal assemblages from
dated sediment cores and that they can be used, based on quan-
titative ecological considerations, to characterise local changes in
EcoQS from pre-impact to present times. For example, by com-
paring the “background” fossil foraminiferal assemblages to the
modern living foraminiferal assemblages at the same site, it would
be possible to determine if a site is naturally anoxic or has become
anoxic with recent human influence. This kind of temporal, in situ
monitoring is not possible with soft-bottom sediment macrofauna
because they do not leave abundant or, for most species, any fos-
sil records. Benthic foraminifera may thus provide a powerful tool
for defining habitat-specific, in situ reference conditions for soft-
bottom coastal and transitional waters. At the moment, there is
no consensus as to which sampling and laboratory methodolo-
gies (e.g. sieving mesh size and use of dead or living assemblages)
are optimal for environmental monitoring. This hampers both
67
comparisons between studies and proposals for using foraminifera
as a tool for stakeholders. However, currently, an expert group, the
FOraminiferal BIo-MOnitoring (FOBIMO)-group (Schönfeld et al.,
2011), is working to get consensus on a standardised methodology.
The present study aims to address parts of these problems by
focusing on aspects concerning species assemblages and diversity
indices which are commonly used for defining EcoQS in environ-
mental monitoring studies. Internationally, there is an increasing
demand for quicker and more cost-efficient methods (as compared
to the traditional ones) to describe ecological status in marine
waters. Consequently, an important objective is to optimise the
methods to find the best compromise between effort and accuracy.
Key questions for defining the EcoQS using benthic foraminiferal
species diversity include the following. What thickness of sediment
should be analysed? Should the >63 ␮m-fraction be used or will the
>125 ␮m-fraction provide a similar result? Is significant informa-
tion lost if only fossilisable species are considered? Do living and
dead assemblages reflect similar EcoQS? How many replicates are
required?
The study sites are silled basins along the Norwegian coast of the
Skagerrak (NE North Sea) with relatively stable salinity and temper-
ature conditions but otherwise different environmental properties
(primarily bottom water dissolved oxygen concentration). To eval-
uate the potential of benthic foraminifera as a bio-assessment
tool, EcoQS of sampling sites are determined using living (stained)
benthic foraminifera data compared with bottom water dissolved
oxygen concentration. Dissolved oxygen is chosen as it is one of
the most important environmental gradients in these systems and
because it is frequently impacted by human activity. The study
is part of a comprehensive project (PES) which includes both
foraminifera and macrofauna collected at the same site at the same
time. Macrofauna results and comparisons between foraminifera
and macrofauna along the same environmental gradients will be
presented in forthcoming papers.
2. Material and methods
2.1. Study area and sampling sites
In August 2008, 27 stations in silled basins along the Norwegian
Skagerrak coast, NE North Sea (Fig. 1), were sampled for biologi-
cal and environmental analyses. Stations were selected to provide
an oxygen gradient with stable temperature (5–6 ◦ C; occasionally
8–9 ◦ C at shallow, 40–50 m, depth) and salinity (33–34) conditions
(see details in Table 1) using information from previous studies
(Buhl-Mortensen et al., 2009 and unpublished data from the Nor-
wegian Institute for Water Research (NIVA) and Institute of Marine
Research (IMR)). The study basins commonly experience deep-
water renewals during winter.
2.2. Field sampling
At each station, sediment samples for biogeochemical, sedi-
mentological, benthic foraminiferal, and macrofaunal analyses, and
bottom water samples for dissolved oxygen analyses, were col-
lected. Four sediment cores (8 cm diameter) were collected at each
of the 27 stations with a Gemini gravity corer (a modified version
of the Niemistö corer (Niemistö, 1974); three replicate cores were
used for benthic foraminifera and one for organic carbon and total
nitrogen analyses. Once on deck, the bottom water from just above
the sediment-water interface in two cores was immediately trans-
ferred to Winkler bottles, sealed, and kept dark and cold (∼7 ◦ C) for
subsequent dissolved oxygen analysis. All sediment cores were sec-
tioned on board and, for the present study, the top 0–1 and 1–2 cm
slices were analysed. Sediment samples for total organic carbon and
68 V.M.P. Bouchet et al. / Ecological Indicators 23 (2012) 66–75
Fig. 1. Sampling stations along the Norwegian Skagerrak coast.
total nitrogen analyses were frozen immediately after sectioning.
Foraminiferal samples were preserved in rose Bengal-stained 70%
ethanol (1 g L−1 ) to avoid protoplasm degradation and to distin-
guish living (stained) from dead specimens (discussion in Murray
and Bowser, 2000). At each station, one sub-sample of the top 1 cm
from a Van Veen grab was collected for grain size analyses.
2.3. Laboratory analyses
Bottom-water dissolved oxygen concentrations were analysed
using Winkler titration. Sediments were freeze-dried prior to sedi-
mentological and geochemical analyses. For grain size analyses, the
dried sediment was weighed, soaked in tap water, and washed on a
63 ␮m Endecote-sieve; the >63 ␮m fraction was dried and weighed
and the <63 ␮m-fraction was calculated based on the dry weights.
Total organic carbon and total nitrogen were analysed using a CHN
analyser (Carlo Erba Elemental Analyzer 1106).
Foraminiferal samples were washed through 500 and 63 ␮m
mesh sieves, and the 63–500 ␮m fraction was split using a modified
Elmgren wet splitter (Elmgren, 1973). One eighth of each sample
was re-sieved and all live (stained) foraminifera in the 63–125 and
125–500 ␮m fractions were identified to species level and counted
in the wet state. The number of individuals >500 ␮m relative to
smaller ones was trivial (<0.1%) so including them would not influ-
ence the results. The samples were analysed wet rather than dry as
this allows preservation of all species (including fragile organic-
walled and loosely cemented agglutinated foraminifera). It also
makes it easier to discriminate stained from unstained specimens.
In this study, the wet-picked, >63 ␮m, living (stained) foraminifera
in the surface 0–2 cm sediment, is called the “complete living
assemblage”. Dead assemblage data were obtained by merging the
counted 0–1 and 1–2 cm sub-samples and re-washing them on a
63-␮m sieve to remove surplus stain. They were then dried at 40 ◦ C
and about 300 dead (unstained) tests were dry-picked, mounted on
microslides, identified and counted. Distinctions between “fossil-
isable” and “non-fossilisable” foraminiferal species are based on
the species’ presence in dried sediment samples from >10 cm core
depth in one or more of 305 samples from the Norwegian Skagerrak
coast (Alve, 1991b, 1996, 2000; Alve et al., 2009).
2.4. Data analysis
2.4.1. Diversity indices
The diversity indices Shannon-Wiener index (H , log2 ) (Shannon
and Weaver, 1963) and Hurlbert index (ES(100) ) (Hurlbert, 1971) are
used by the Norwegian Pollution Control Authority as metrics to
characterise EcoQS along the Norwegian coast. Since H and ES100
are highly correlated in our data (r2 = 0.97, p < 0.001), we concen-
trate on H .
H is biased when there are unobserved species in the commu-
nity, a common problem with under-sampling (Chao and Shen,
2003). Chao and Shen (2003) introduce a bias-corrected ver-
sion of Shannon’s index (H bc ), which has little bias (Beck and
Schwanghart, 2010), and which we use in this paper. Shannon’s
index is an entropy rather than a diversity. The entropy gives the
average uncertainty of the identity of an individual picked from
the community, not the number of species in the community (e.g.
Hayek and Buzas, 1997; Jost, 2006). It can be converted to true
diversity, the effective number of species, with the exponential
function (N1 = exp(H bc ), Hill, 1973). Exp(H bc ) gives the number
of species that would, if each were equally common, produce the
same H bc as the sample. Given a community sample of three
species represented by 100, 50, and 100 individuals, H bc = 1.05 and
exp (H bc ) = 2.87. Thus 2.87 species of equal abundance gives an H bc
of 1.05. To determine how many foraminifera need to be counted
to get a stable estimate of diversity, we resampled each sample
100 times, with replacement, to different sized subsamples and cal-
culated the coefficient of error of the effective number of species.
Correlations between the bias-corrected exponent of Shannon of
different fractions of the foraminiferal community, as well as of
their microhabitat and the environmental variables, are calculated.
All data represent the pooled counts from three replicates per sta-
tion (rather than the average).
Benthic foraminiferal data are used to evaluate EcoQS at the
sampling stations. As no criteria exist for determining EcoQS using
benthic foraminifera, it has to be defined. In this study, we set
the highest expected exp(H bc ) to 25 effective species (0–2 cm,
>63 ␮m, wet-picked, living assemblages, Table 2), a value slightly
higher than the highest observed value of 22 effective species. Any
 V.M.P. Bouchet et al. / Ecological Indicators 23 (2012) 66–75 69

Table 1
Characteristics of sampling stations (nd: no data): water depth (m), sill depth (m, Norwegian Hydrographic Service), bottom-water dissolved oxygen (minimum value over
the 2006–2008 period, Min[O2 ]2 years , mL O2 L−1 ), bottom-water dissolved O2 -concentration at the time of sampling ([O2 ]tos , mL O2 L−1 ), sediment grain size (%, >63 ␮m),
total organic carbon (TOC, %), total nitrogen (TN, %), C/N ratio, living and dead foraminiferal abundances (0–2 cm, >63 ␮m, ind. cm−3 ), diversity of living foraminiferal
assemblages (H , ES(100) and exp(H bc ) for complete living assemblages and exp(H bc ) for 0–1, >125 ␮m, living fossilisable assemblages) and EcoQS. Colours reflect EcoQS;
see Table 2 for location of areas, see Fig. 1.

assemblages with more than 25 effective species would be set to
25. We use zero effective species to represent an azoic sample.
Within this range (0–25 effective species) we follow Rosenberg
et al. (2004) and divide this range into the five equal-sized
ecological status classes (unacceptable statuses: “bad”, “poor”,
“moderate”, acceptable statuses: “good” and “high”) required by
the WFD classification (Table 2). The top two, and the bottom
three, classes can be combined to give the MSFD classes “good”
and “bad”. Accuracy of a less time-consuming method using diver-
sity derived from >125 ␮m, fossilisable, living assemblages in the
surface 0–1 cm of sediment has been tested. Criteria are presented
in Table 2.

Table 2
Criteria for determining EcoQS using living benthic foraminifera.

EcoQS and associated colour code Bad Poor Moderate Good High

EcoQS derived from 0-2 cm, >63 µm,

<5 5-10 10-15 15-20 >20
living wet-picked assemblages

EcoQS derived from 0-1 cm, >125 µm,

<2.5 2.5-5 5-7.5 7.5-10 >10
living dry-picked assemblages
70
Standard Deviation of effective no. of species
4
3
2
0 1
5 10 15
Mean effective number of species
Fig. 2. Relationship between the standard deviation and mean of the effective num-
ber of living (stained) species (complete living assemblages) in the three replicates.
To estimate the probability of assigning an assemblage to an
incorrect EcoQS class, and how this varies with the number of repli-
cates, we use the methods developed by Kelly et al. (2009). We find
the relationship between the standard deviation and mean of the
effective number of species in the three replicates (Fig. 2). Using
this relationship, we estimate the standard error expected for any
value of the mean given one, three or six replicates, and there-
fore the distribution of uncertainty for each mean with the given
number of replicates. We can then estimate the probability that
any sample will be misclassified. See Kelly et al. (2009) for further
details.
2.4.2. Community analysis
Redundancy analysis was used to find the proportion of
the variance of the different fractions of the square-root
transformed community data explained by the environmental vari-
ables. Procrustes analysis (Peres-Neto and Jackson, 2001) was
used to compare principal component analyses of the differ-
ent fractions of the foraminiferal community. We report the
m2 statistics, which is analogous with the r2 of a correla-
tion.
All calculations were performed using the statistical language R
version 2.11.1 (R Development Core Team, 2010). Ordinations were
run with the vegan library version 1.17-2 (Oksanen et al., 2010).
Bias-corrected Shannon Index was calculated with the Entropy
library version 1.1.5 (Hausser and Strimmer, 2010).
3. Results
The absolute abundance of living (stained) benthic foraminifera
(>63 ␮m; all taxa included) in the surface 0–2 cm of sediment
ranged from 6 to 201 ind. cm−3 (wet sediment) (Table 1) with an
average of 58 ± 41 ind. cm−3 . Of these individuals, 76% occurred in
the top 0–1 cm. In total, 116 living (stained) benthic foraminiferal
species were recorded of which 88 (or 76%) are potentially fossil-
isable and these species account for 89% of the living foraminiferal
abundance (>63 ␮m, 0–1 cm). For the dead assemblages (0–2 cm,
>63 ␮m), the absolute abundance ranged from 3 to 514 ind. cm−3
(Table 1).
V.M.P. Bouchet et al. / Ecological Indicators 23 (2012) 66–75
40
Coefficient of error
30
20
10
0 100 200 300 400 500
20
Number individuals
Fig. 3. Coefficient of error of the effective number of species against the number of
individuals (complete living assemblages). Each line represents one sample, resam-
pled, with replacement, to different sized subsamples. Values are the result of 100
trials.
3.1. Living foraminiferal diversity and bottom-water dissolved
oxygen
The coefficient of error of exp(H bc ) is quite high if less than 50
individuals are counted and identified (Fig. 3). There is a progressive
decrease in the error until about 250 individuals have been counted,
after which the error is low and relatively stable.
Bottom-water oxygen concentration at the time of sampling
([O2 ]tos ) is a good predictor (Table 3) of the effective number
of species, exp(H bc ), whereas grain size, TOC, TN and C/N are
weaker. Following Diaz and Rosenberg’s (2008) definition for oxy-
gen conditions, anoxia (0.0–0.5 mL O2 L−1 ) characterizes stations
R180, R160, R140, R120 and 102, hypoxia (0.5–2.0 mL O2 L−1 ) sta-
tions F90, R100, F70 and R80 ([O2 ]tos , Table 1). All the other
stations have normoxic (>2.0 mL O2 L−1 ) conditions. There is a
strong correlation between exp(H bc ) between the complete living
foraminiferal assemblages diversity (0–2 cm, >63 ␮m) and [O2 ]tos
(r = 0.79, Table 3). Additionally, [O2 ]tos explains a large proportion of
the variance in the complete living foraminiferal community data.
3.2. Comparison of foraminiferal sampling and processing
methods
The confidence with which samples can be assigned to each
status class can be represented as a bell-shaped curve with the
maximum at the centre of the relevant status class (Fig. 4A). The
tails of these curves overlap, so an observed diversity could poten-
tially belong to several EcoQS classes. As an example, if the diversity
was 17.5, the confidence of the condition of the station being “good
status” is about 65% (based on a single replicate), and there is an
approximately 18% chance that the “true” condition is “high” or
“moderate”. If there are 3 replicates, then there is about 90% con-
fidence that the true condition is “good”, and the chance of the
true condition being “high” or “moderate” being approximately 7%
(Fig. 4B). The risk of misclassification, i.e. of placing a site in any sta-
tus class other than the true one, decreases with distance from class
boundaries, wfith the lowest risk of misclassification occurring at
the centre of a status class. At the class boundaries, it is equally
likely that the true EcoQS is in either the lower or the higher status
class. The risk of misclassification decreases as the number of repli-
cates increases (Fig. 4C). In the present example, there is a risk of
 V.M.P. Bouchet et al. / Ecological Indicators 23 (2012) 66–75 71

Bad Poor Moderate Good High

1.0
A
Confidence of Class, n=1
0.8
0.6
0.4
0.2
1.0 0.0

B
Confidence of Class, n=3
0.8
0.6
0.4
0.2
0.0

C
0.5
Missclassification rate
0.4
0.3
0.2
0.1
0.0

5 10 15 20 25
exp(H′bc)

Fig. 4. Confidence of a correct ecological status class prediction for sampling stations using complete living assemblages, based on A: a single replicate (dash-dot: bad EcoQS,
solid: poor, dash: moderate, dot: good, dash-dot: high) B: 3 replicates (dash-dot: bad EcoQS, solid: poor, dash: moderate, dot: good, dash-dot: high). C: risk of face-value
misclassification for sampling stations based on different numbers of replicates (solid: 1 replicate, dash: 3 replicates and dot: 6 replicates).

misclassification of about 35% for a diversity of 17.5, taking a single
replicate. The risk drops to approximately 10% with 3 replicates,
and is negligible with 6 replicates.
The correlation between living, >63 ␮m foraminiferal exp H bc
in the 0–1 cm and [O2 ]tos is stronger (r = 0.77, Table 3) than for that
of the 1–2 cm assemblages (r = 0.67, Table 3). exp(H bc ) values of the
assemblages in the different size fractions are also correlated with
[O2 ]tos (r = 0.68 for the >125 ␮m and r = 0.68 for the 63–125 ␮m,
Table 3). Similar results are achieved using community composition
instead of diversity.
The relationship between exp(H bc ) and [O2 ]tos is stronger for
the fossilisable assemblages than for the non-fossilisable ones
(r = 0.79 and r = 0.42, respectively, Table 3). Correlation between the
>125 ␮m fraction of live fossilisable assemblages in the 0–1 cm and
[O2 ]tos is slightly weaker than for the complete living assemblages
(r = 0.69 and r = 0.79, respectively, Table 3).
Correlation between exp(H bc ) and [O2 ]tos is stronger for the
complete living assemblages than for the dead assemblages (r = 0.79
and r = 0.54, respectively, Table 3). The diversity in the complete
living assemblages is better correlated to the diversity in the liv-
ing large fossilisable assemblages than to the diversity in the dead
assemblages (r = 0.86 and r = 0.64, respectively, Table 3). Commu-
nity data from the complete living assemblages and the dead
assemblages, and the living fossilisable assemblages and the dead
assemblages are also correlated (m2 = 0.59 and m2 = 0.60, respec-
tively, Table 3).
3.3. EcoQS based on foraminifera data
Exp(H bc ) values vary from 1.2 at stations R180 and 6
to 22.0 at station IH30 (complete living assemblage, 0–2 cm,
>63 ␮m, Table 1). Ecological quality status of anoxic stations
is “bad” (Fig. 5). At hypoxic sites, stations are all ranked as
“bad” EcoQS except station F70 which is “poor”. At normoxic
stations, EcoQS vary from “bad” at stations 6, G60, G69 and G50 to
high at stations 106 and IH30. Over the last 2 years, stations 6, G60,
G69 and G50 experienced periods of severe hypoxia (Min[O2 ]2 years ,
Table 1).
In this study, EcoQS derived from the diversity of the complete
living assemblages are “non acceptable” (bad-poor-moderate) at
19 stations, and “acceptable” (good-high) at 8 stations (Table 1).
Using the diversity of the 0–1 cm, >125 ␮m, dry-picked, living
72 V.M.P. Bouchet et al. / Ecological Indicators 23 (2012) 66–75

Table 3
Correlation of diversity (exp(H bc )) against [O2 ]tos for different live fractions in different sediment layers and for dead foraminifera. Correlation of diversity of the different
subsets against diversity in the complete living assemblage. Proportion of the square rooted relative abundances explained by [O2 ]tos in an RDA. Procrustes rotation m2 of PCA
of complete assemblage against different fractions. Complete assemblage = all live (stained) foraminifera >63 ␮m in size. All correlations are significant at p < 0.05 (numbers
in brackets are correlation or m2 of dead foraminifera against fossilisable living foraminifera).

Correlation Correlation diversity complete Proportion community Procrustes m2 against

diversity vs [O2 ]tos living assemblage vs diversity variance explained by [O2 ]tos complete living assemblage
of subsets

Complete living assemblage 0.79 1.00 0.28 1.00

(specimens > 63 ␮m, 0–2 cm)
Small living 0.68 0.91 0.21 0.81
(63 ␮m < Specimens < 125 ␮m,
0–2 cm)
Large living 0.68 0.87 0.27 0.97
(specimens > 125 ␮m, 0–2 cm)
Shallow living 0.77 0.96 0.31 0.98
(specimens > 63 ␮m, 0–1 cm)
Deep living 0.67 0.85 0.19 0.76
(specimens > 63 ␮m, 1–2 cm)
Fossilisable living 0.79 0.98 0.28 0.99
(specimens > 63 ␮m, 0–2 cm)
Not fossilisable living 0.42 0.60 0.11 0.45
(specimens > 63 ␮m, 0–2 cm)
Large fossilisable living 0.69 0.86 0.29 0.96
(specimens > 125 ␮m, 0–1 cm)
Dead (specimens > 63 ␮m, 0.54 0.64 (0.69) 0.14 0.59 (0.60)
0–2 cm))

25 Anoxic Hypoxic Normoxic studies (see points to consider in Murray, 2006). In scientific stud-
High
ies, the choice of methods depends on the nature of the study area
IH30
20 106 and the ecological questions in focus. However, to promote ben-
KØ
71 thic foraminifera as an efficient environmental monitoring tool to
IH45 Good
F50 G40 assess EcoQS, there is a need to test and identify methodologies
exp (H'bc)

15 50
Moderate
which are optimal both from a scientific and a practical point of
R60 F30
200 view. Key issues addressed in the present study include:
10 IH60

F70 Poor
- Replication.
5
R140
102 R100
G50
- Thickness of sediment layer analysed.
R80 Bad
R160 R120 G69 - Choice of size fraction.
R180 F90 6 G60
0
0 1 2 3 4 5 - Analyses of wet (i.e. complete live (stained) assemblage) vs dry
Bottom-water dissolved oxygen ([O2]tos, mL O2.L-1) samples (i.e. fossilisable species).
- Analyses of live or dead assemblages.
Fig. 5. exp(H bc ), of complete living (stained) assemblage (0–2 cm, >63 ␮m, pooled
replicates) and corresponding EcoQ, against [O2 ]tos . Threshold between anoxic, 4.1.1. Replication
hypoxic and normoxic bottom-water conditions are from Diaz and Rosenberg
Patchiness in distribution patterns of benthic foraminifera may
(2008).
occur both on 10-cm and 1-m sample spacing (e.g. Boltovskoy and
Lena, 1969; Buzas et al., 2002). It is thus an issue to be considered
assemblages, EcoQS are “non acceptable” at 15 stations, and in environmental studies. Replicate sampling aims to minimize this
“acceptable at 12 stations. Disagreements between the 2 methods problem. However, for historical reasons, the use of replicates has
occur at stations 200, F70, R60 and KDR. not been common practice in benthic foraminiferal studies and in
most of those which have included replicates, abundance patterns
4. Discussion rather than diversity have been considered.
In this study, the risk of misclassification of the EcoQS is high if
By applying the criteria for EcoQS set by the Norwegian Pol- data are based on only one sample per station (Fig. 4). The degree
lution Control Authority to micropaleontological and geochemical of confidence is higher if three replicates per station are taken. Risk
data from near-shore sediment cores, Alve et al. (2009) showed of misclassification if six replicates are considered is lower, but the
the potential of using fossil (dead) benthic foraminifera to evalu- improvement observed from three replicates is not as large as the
ate temporal changes in EcoQS from “reference” (background) to improvement from one to three replicates. Results presented here
present-day conditions. Correlation between bottom-water oxy- provide a basis for sampling methods using benthic foraminifera
gen concentration and faunal diversity in the present study takes in environmental monitoring, and suggest that three replicates are
this a substantial step further. To our knowledge, this is the first sufficient to determine a reliable EcoQS.
work using benthic foraminifera to show quantitative relationships
between environmental variables and diversity in coastal areas, and 4.1.2. Thickness of sediment layer analysed
to evaluate EcoQS. Benthic foraminifera may live several decimeters below the
sediment-water interface (see review in Bouchet et al., 2009); how
4.1. Methodological considerations deep they live depends on local geochemical conditions and food
availability (e.g. Corliss and Emerson, 1990; Gross, 2000; Jorissen
In the literature, there is an obvious lack of standardisation of et al., 1995; Linke and Lutze, 1993). To obtain the most represen-
both sampling and processing methods in foraminiferal ecological tative view of the live (stained) benthic foraminiferal diversity at
 V.M.P. Bouchet et al. / Ecological Indicators 23 (2012) 66–75
a site, each sample should ideally include the entire sediment col-
umn inhabited by the foraminifera. Obviously, this is impractical
and most studies base diversity calculations on assemblages occur-
ring in the surface 0–1 or 0–2 cm of sediment. In this study, we
have shown that a substantial proportion of the individuals live
in the sub-surface 1–2 cm of sediment. Still, although the diver-
sity (>63 ␮m, all living taxa) of both the 0–1 cm and 1–2 cm are
significantly correlated with bottom water [O2 ]tos , the correlation
is better for the surface 0–1 cm. This should not be surprising as
samples deeper than 1 cm reflect [O2 ] in the sediment pore water
rather than bottom water [O2 ]. For the present investigation areas
it follows that using living benthic foraminiferal diversity indices,
simply representing the 0–1 cm of sediment gives an accurate eval-
uation of the ecological quality of a study site. However, it also
indicates that using the 0–2 cm sediment level gives comparable
results if this is preferred for other reasons.
4.1.3. Choice of size fraction
The choice of size fraction (e.g. >63, >125, >150 ␮m) in
foraminiferal ecological studies has generated a number of pub-
lications. For example, Schröder et al. (1987), in their study of total
assemblages (living + dead), showed that focusing on the coarser
(>125 ␮m) rather than the finer (>63 ␮m) fraction may lead to a
loss of indicator species and differences in species abundances.
The impact of the choice of size fraction on diversity indices is
poorly known. In our study, the diversity and assemblage compo-
sition of the >63 ␮m and the >125 ␮m fraction assemblages are
significantly correlated, and the diversity of both fractions are cor-
related to bottom water [O2 ]tos , with the >63 ␮m fraction giving
a slightly better correlation. However, analysing just the coarse
fraction is less time consuming, and requires less taxonomic skill
than analysing the complete assemblage or the fine fraction. For
many species, accurate identification of juveniles to species level
may be difficult. This is important, as in addition to its accuracy,
a bio-indicator should ideally be fast and easy to apply if it is
to be widely adopted. Since the information lost by just using
the coarse (>125 ␮m) fraction is small, this is an adequate choice
for calculation of diversity and assemblage indices for monitoring
EcoQS.
4.1.4. Analyses of wet vs dry samples
To obtain the most complete picture possible of the
foraminiferal assemblages (i.e. including soft- as well as hard-
shelled forms) in the study areas, all living foraminiferal samples
were wet-counted. However, we did not obtain a substantially bet-
ter correlation with bottom water [O2 ]tos by including the fragile
forms in the computation of the diversity indices. The fact that the
fossilisable fraction of the living assemblages correlates better with
the bottom water [O2 ]tos than the fragile, non-fossilisable compo-
nent (Table 3) has important implications for the use of benthic
foraminifera as palaeoecological status (PalaeoEcoQS) indicators.
It implies that, in the present study, the signal contained in the
fossilisable assemblage diversity provides an accurate picture of
the present-day environmental conditions. This means that even
if most of the fragile living forms will not be preserved in the fos-
sil assemblages, the signal given by fossil assemblages would give
a good picture of the palaeoenvironmental conditions. This shows
that in the present study areas, using the dry-picking method is
adequate for reliably defining EcoQS based on benthic foraminiferal
diversity. Dry-picking is advantageous because it is quicker. On the
other hand, analysing the assemblages in the wet state makes it
easier to distinguish between stained and unstained specimens,
particularly for some agglutinated and opaque miliolid forms. To
some extent, this can be compensated for by wetting or crushing
the tests.
73
4.1.5. Analyses of living or dead assemblages
In the present study, it is not surprising that the dead
foraminiferal assemblages do not correlate as well as the living
ones with the bottom water [O2 ]tos . Dead assemblages repre-
sent the mixing of tests from a succession of previously living
assemblages modified by taphonomic processes (Murray, 2000),
i.e. reflecting time-integrated environmental conditions commonly
spanning several years. The time-period represented depends on
the sedimentation accumulation rate. On the other hand, liv-
ing assemblages reflect the environmental conditions which have
operated at a site during the life-span of the organisms present,
i.e. the EcoQS at the time of sampling. Unpublished dating-results
of sediment cores from 5 of the investigated sites, show that the
dead assemblages in the surface 0–2 cm represent tests (shells)
accumulated during time periods of 2 to >10 years. Consequently,
the living and dead assemblages reflect two principally differ-
ent processes and, thereby, different aspects of the environmental
conditions at a site. This is why the use of total (live + dead)
assemblages, as still recommended by some authors, is difficult
to interpret meaningfully and should be avoided (see also dis-
cussion in Murray, 2000). Optimal environmental information is
gained by analysing and treating the two separately. As illus-
trated by the better correlation with the bottom water [O2 ]tos ,
the living foraminiferal assemblages give a more accurate eval-
uation of EcoQS at the time of sampling whereas the dead ones
reflect the longer time trends in EcoQS. Additionally, in the present
study areas, picking living assemblages is not more time consum-
ing as their abundance, relative to dead individuals is quite high
(Table 1).
4.2. Evaluate EcoQS using living benthic foraminifera
This is the first study defining criteria to determine EcoQS using
living benthic foraminifera for environmental monitoring. At our
study sites, [O2 ]tos reflects a disturbance gradient. Living ben-
thic foraminiferal assemblages diversity (the effective number of
species of complete living assemblages) is significantly correlated
with [O2 ]tos (Table 3). Improved EcoQS as reflected by foraminiferal
diversity is observed along with increasing oxygen concentra-
tions. Disturbed stations, i.e. here exposed to anoxic conditions,
are classified as reflecting “bad” EcoQS based on foraminiferal
diversity (Fig. 5). On the other hand, well-oxygenated stations are
ranked from “moderate” to “high” EcoQS using the diversity index
exp(H bc ). We obtained a clear pattern from bad to high EcoQS
using living benthic foraminifera, and this is in accordance with the
bottom-water oxygen gradient ([O2 ]tos ). Indeed, different studies
reported drops in benthic foraminiferal diversity in response to a
gradient of disturbance (Alve et al., 2009; Elshanawany et al., 2011;
Mojtahid et al., 2008). Exceptions are reported at stations in Top-
dalfjord (6) and Groosefjord (G50, G60 and G69) with normoxic
conditions at the time of sampling but low diversity, and therefore
“bad” EcoQS. Stations 6, G50, G60 and G69 had experienced peri-
ods of severe hypoxia over the past 2 years. The minimum value
during the recent past at each site captures the extreme condi-
tions which the foraminifera have experienced. They have higher
TOC concentrations than stations with similar bottom-water oxy-
gen concentration at the time of sampling. Buhl-Mortensen et al.
(2009) reported that fjords in Southern Norway with low histori-
cal oxygen level also had in general a high carbon content in the
surface sediment. This may indicate that although bottom-water
[O2 ]tos is a good predictor of the foraminiferal diversity (and bet-
ter than TOC), [O2 ]tos in the near-bottom water is not the variable
that the infaunal foraminifera are actually responding to. Sediment
pore-water conditions are more direct drivers of foraminiferal com-
munity composition, but will in fjord systems, usually be correlated
with the properties of the bottom water.
74 V.M.P. Bouchet et al. / Ecological Indicators 23 (2012) 66–75
In this study, we are suggesting that using the 0–1 cm, >125 ␮m,
living dry-picked assemblages would give an accurate evaluation
of the EcoQS. Using this suggested method means recording fewer
species of the living assemblages, criteria has thus been adjusted
(Table 2), as diversity values cannot be as high as for the complete
living assemblages. We manage to highlight the gradient of dis-
turbance using this method. Comparing EcoQS derived from the
complete living assemblages and from this less-time consuming
method, we obtain almost a full agreement in the classification
(unacceptable or acceptable), with disagreements at only 4 stations
(Table 1).
Consequently, diversity-based indices derived from benthic liv-
ing foraminifera seem to be an adequate method to highlight
a gradient of disturbances, and to determine EcoQS. However,
it needs to be adjusted and further tested against natural or
human-induced disturbances other than oxygen and in different
environments.
4.3. Determine reference conditions using fossil benthic
foraminifera
Since the assessment of ecological quality is based on the
extent of deviation from reference conditions (WFD, 2000/60/EC),
it is crucial to know the “true” reference conditions to obtain a
valid classification. Currently, environmental monitoring studies
use samples from similar and supposedly non-impacted environ-
ments as reference conditions or use expert judgment to define
it (e.g. Bigot et al., 2008; Borja and Tunberg, 2011; Borja et al.,
2012; Bouchet and Sauriau, 2008; Muxika et al., 2007). Use of
historical data, as suggested by the WFD, would provide an objec-
tive way of defining reference conditions but such data are rarely
available. Alternatively, Alve et al. (2009) suggested that benthic
foraminifera have a good potential to be used to determine refer-
ence conditions using the fossil record, provided possible impacts
of taphonomic processes are considered. To what extent the sig-
nal contained in the living assemblages is preserved in the dead
ones remains uncertain. In this study, diversity of living fossilisable
and dead assemblages are correlated. This indicates that in areas
with minimal tidal influence (minimal transport of tests) and sed-
iment accumulation rates of one or a few mm year−1 (common in
S Norwegian fjords), the diversity of dead assemblages can be used
as a proxy for the diversity of past (late summer) living commu-
nities. This correlation suggests that the environmental signal in
the living foraminifera is well preserved in the fossil record. Hence,
analyses of fossil benthic foraminifera allow for reconstruction of
PalaeoEcoQS back to pre-impacted times (i.e. reference conditions).
The detection of anthropogenic signals against background envi-
ronmental noise is an issue in environmental monitoring studies
(Elliott and Quintino, 2007), and requires long term data, e.g. from
palaeoecological work. It is very difficult to distinguish between a
faunal assemblage with a low diversity due to a simply naturally
stressed environment and one with low diversity induced by an
anthropogenic stress. Looking at the PalaeoEcoQS reconstruction,
it would be possible to determine whether or not the site has always
been stressful for the benthic fauna.
The preservation of benthic foraminifera in the fossil record
is a great asset compared to the non-preservation of benthic
macrofauna, the traditional monitoring tool. The present results
suggest that diversity-based indices derived from living ben-
thic foraminifera represent a promising tool to assess EcoQS
not only under present-day conditions but also in the past. It
is also mentioned in the WFD that reference conditions are
type-specific, so they are different for different types of coastal
waters so as to take into account the broad diversity of ecolog-
ical regions in Europe. It might actually be more appropriate to
use “site-specific” boundaries between EcoQS classes instead of
“type-specific” (Krause-Jensen et al., 2005) to decrease the risk
of misinterpreting EcoQS. Criteria established in this study would
then need to be adjusted at each site for a more accurate use of
benthic foraminifera diversity as a bio-monitoring tool.
5. Conclusion
In this study, we show that consistent evaluation of ecological
quality status (EcoQS) using the effective number of species derived
from living (stained) benthic foraminifera can be obtained by sam-
pling the 0–1 cm of sediment, pick about 250 living individuals
of fossilisable species (i.e. dry-picked) from the >125 ␮m-fraction
of each of three replicates per station. Sampling and process-
ing benthic foraminiferal samples using this method provide a
quick and efficient method to evaluate EcoQS. Correlation between
living foraminiferal diversity and [O2 ]tos , as well as correlation
between complete community data (here surface 0–2 cm, >63 ␮m,
living (stained) foraminifera) and [O2 ]tos , suggest that benthic
foraminifera can be an efficient bio-monitoring tool to evaluate
EcoQS. At the moment, researchers are having problems finding and
corroborating “true” reference conditions in present-day ecosys-
tems. Since foraminifera are preserved in the fossil record, it would
be possible to use this biological element to evaluate PalaeoE-
coQS and estimate “true” reference conditions. The use of benthic
foraminifera as a bio-monitoring tool would be a benefit for envi-
ronmental monitoring programs worldwide. It nevertheless needs
to be further assessed and tested in other ecosystems and against
different types of pollution pressures.
Acknowledgements
We are grateful to the crew of the R/V Trygve Braarud; S. Holm,
J. Sundøy and T.E. Baade. Thanks are also due to M. Hollerbach,
E. Oug and J. Håvardstun for assistance during fieldwork, and Y.
Descatoire for graphics. We thank Dr. B. Hayward and an anony-
mous reviewer for constructive comments. We also want to thank
J. Magnusson and J. Aure for providing Min[O2 ]2 years data, and S.
Hess, N. Reuss, C. Duffield and J.W. Murray for useful discussions
and constructive comments on the manuscript. This study was sup-
ported by the Norwegian Research Council-funded project PES (no.
184870) “Paleoecological reconstructions of marine soft-bottom
Ecologic Status and in situ reference conditions: calibrating benthic
foraminifera with macrofauna and hydrographic data”. V.M.P.B. is
supported through the PES project by a post-doctoral fellowship.
References
Alve, E., 1991a. Benthic foraminifera in sediment cores reflecting heavy metal pol-
lution in Sorfjord, Western Norway. Journal of Foraminiferal Research 21, 1–19.
Alve, E., 1991b. Foraminifera, climatic change and pollution: a study of Late Holocene
sediments in Drammensfjord, SE Norway. The Holocene 1, 243–261.
Alve, E., 1995. Benthic foraminiferal responses to estuarine pollution: a review.
Journal of Foraminiferal Research 25, 190–203.
Alve, E., 1996. Benthic foraminiferal evidence of environmental change in the
Skagerrak over the past six decades. NGU Bulletin 430, 85–93.
Alve, E., 2000. Environmental stratigraphy: a case study reconstructing bottom
water oxygen conditions in Frierfjord, Norway, over the past five centuries. In:
Martin, R. (Ed.), Environmental Micropaleontology. Kluwer Academic/Plenum
Publishers, New York, pp. 323–350.
Alve, E., Lepland, A., Magnusson, J., Backer-Owe, K., 2009. Monitoring strategies
for re-establishment of ecological reference conditions: possibilities and lim-
itations. Marine Pollution Bulletin 59, 297–310.
Armynot du Châtelet, E., Debenay, J.-P., Soulard, R., 2004. Foraminiferal proxies for
pollution monitoring in moderately polluted harbors. Environmental Pollution
127, 27–40.
Ballesteros, E., Torras, X., Pinedo, S., García, M., Mangialajo, L., de Torres, M., 2007.
A new methodology based on littoral community cartography dominated by
macroalgae for the implementation of the European Water Framework Direc-
tive. Marine Pollution Bulletin 55, 172–180.
Beck, J., Schwanghart, W., 2010. Comparing measures of species diversity from
incomplete inventories: an update. Methods in Ecology and Evolution 1, 38–44.
 V.M.P. Bouchet et al. / Ecological Indicators 23 (2012) 66–75
Bigot, L., Gremare, A., Amouroux, J.-M., Frouin, P., Maire, O., Gaertner, J.-C., 2008.
Assessment of the ecological quality status of soft-bottoms in Reunion Island
(tropical Southwest Indian Ocean). Marine Pollution Bulletin 56, 704–722.
Blanchet, H., Lavesque, N., Ruellet, T., Dauvin, J.-C., Sauriau, P.-G., Desroy, N.,
Desclaux, C., Leconte, M., Bachelet, G., Janson, A.-L., Bessineton, C., Duhamel,
S., Jourde, J., Mayot, S., Simon, S., de Montaudouin, X., 2008. Use of biotic indices
in semi-enclosed coastal ecosystems and transitional waters habitats – impli-
cations for the implementation of the European water framework directive.
Ecological Indicators 8, 360–372.
Boltovskoy, E., Lena, H., 1969. Microdistribution des foraminifères benthoniques
vivants. Revue de micropaléontologie 12, 177–185.
Borja, A., Tunberg, B.G., 2011. Assessing benthic health in stressed subtropical estu-
aries, eastern Florida, USA using AMBI and M-AMBI. Ecological Indicators 11,
295–303.
Borja, A., Dauer, D.M., Gremare, A., 2012. The importance of settings targets and ref-
erence conditions in assessing marine ecosystem quality. Ecological Indicators
12, 1–7.
Bouchet, V.M.P., Debenay, J.-P., Sauriau, P.-G., Radford-Knoery, J., Soletchnik, P.,
2007. Effects of short-term environmental disturbances on living benthic
foraminifera during the Pacific oyster summer mortality in the Marennes-Oléron
Bay (France). Marine Environmental Research 64, 358–383.
Bouchet, V.M.P., Sauriau, P.-G., 2008. Influence of oyster culture practices and envi-
ronmental conditions on the ecological quality of intertidal mudflats in the
Pertuis Charentais (SW France): a multi-index approach. Marine Pollution Bul-
letin 56, 1892–1912.
Bouchet, V.M.P., Sauriau, P.-G., Debenay, J.-P., Mermillod-Blondin, F., Schmidt, S.,
Amiard, J.-C., Dupas, B., 2009. Influence of the mode of macrofauna-mediated
bioturbation on the vertical distribution of living benthic foramnifera: first
insight from axial tomodensitometry. Journal of Experimental Marine Biology
and Ecology 371, 20–33.
Buhl-Mortensen, L., Oug, E., Aure, J., 2009. The response of hyperbenthos and infauna
to hypoxia in fjords along the Skagerrak: estimating loss of biodiversity due to
eutrophication. In: Moksness, E., Dahl, E., Støttrup, J.G. (Eds.), Integrated Coastal
Zone Management. , pp. 79–96.
Burone, L., Venturini, N., Sprechmann, P., Valente, P., Muniz, P., 2006. Foraminiferal
responses to polluted sediments in the Montevideo coastal zone, Uruguay.
Marine Pollution Bulletin 52, 61–73.
Buzas, M.A., Hayek, L.C., Reed, S.A., Jett, J.A., 2002. Foraminiferal densities over five
years in the Indian River lagoon, Florida: a model of pulsating patches. Journal
of Foraminiferal Research 32, 68–92.
Chao, A., Shen, T.-J., 2003. Nonparametric estimation of Shannon’s index of diver-
sity when there are unseen species in sample. Environmental and Ecological
Statistics 10, 429–443.
Coates, S., Waugh, A., Anwar, A., Robson, M., 2007. Efficacy of a multi-metric fish
index as an analysis tool for the transitional fish component of the Water Frame-
work Directive. Marine Pollution Bulletin 55, 225–240.
Corliss, B., Emerson, S., 1990. Distribution of rose-bengal stained deep-sea benthic
foraminifera from the Nova Scotian continental-margin and Gulf of Maine. Deep-
Sea Research Part A-Oceanographic Research Papers 37, 381–400.
Debenay, J.-P., Fernandez, J.-M., 2009. Benthic foraminifera records of complex
anthropogenic environmental changes combined with geochemical data in
a tropical bay of New Caledonia (SW Pacific). Marine Pollution Bulletin 59,
311–322.
Diaz, R., Rosenberg, R., 2008. Spreading dead zones and consequences for marine
ecosystems. Science 321, 926–929.
Elliott, M., Quintino, V., 2007. The Estuarine Quality Paradox, environmental homeo-
stasis and the difficulty of detecting anthropogenic stress in naturally stressed
areas. Marine Pollution Bulletin 54, 640–645.
Elmgren, R., 1973. Methods of sampling sublittoral soft bottom meiofauna. Oikos
Supplementum 15, 112–120.
Elshanawany, R., Ibrahim, M.I., Milker, Y., Schmiedl, G., Badr, N., Kholeif, S.E.A., Zonn-
eveld, K.A.F., 2011. Anthropogenic impact on benthic foraminifera, Abu-Qir Bay,
Alexandria, Egypt. Journal of Foraminiferal Research 41, 326–348.
Gooday, A.J., 1988. A response by benthic foraminifera to the deposition of phytode-
tritus in the deep sea. Nature 332, 70–73.
Gross, O., 2000. Influence of temperature, oxygen and food availability on the
migrational activity of bathyal benthic foraminifera: evidence by microcosm
experiments. Hydrobiologia 426, 123–137.
Gustafsson, M., Nordberg, K., 2001. Living (stained) benthic foraminiferal response
to primary production and hydrography in the deepest part of the Gullmar Fjord,
Swedish west coast, with comparisons to Höglund’s 1927 material. Journal of
Foraminiferal Research 31, 2–11.
Hausser, J., Strimmer, K., 2010. Entropy: Entropy and Mutual Information Estimation.
Hayek, L.A.C., Buzas, M.A., 1997. Surveying Natural Populations. Columbia University
Press, New York, 563 pp.
75
Hayward, B.W., Grenfell, H.R., Nicholson, K., Parker, R., Wilmhurst, J., Horrocks, M.,
Swales, A., Sabaa, A.T., 2004. Foraminiferal record of human impact on intertidal
estuarine environments in New Zealand’s largest city. Marine Micropaleontol-
ogy 53, 37–66.
Hill, M., 1973. Diversity and evenness: a unifying notation and its consequences.
Ecology 54, 427–432.
Hurlbert, S.H., 1971. The non-concept of species diversity. Ecology 52, 577–586.
Jorissen, F.J., de Stigter, H.C., Widmark, J.G.V., 1995. A conceptual model explaining
benthic foraminiferal microhabitats. Marine Micropaleontology 26, 3–15.
Jost, L., 2006. Entropy and diversity. Oikos 113, 363–375.
Kelly, M., Bennion, H., Burgess, A., Ellis, J., Juggins, S., Guthrie, R., Jamieson, J., Adri-
aenssens, V., Yallop, M., 2009. Uncertainty in ecological status assessments of
lakes and rivers using diatoms. Hydrobiologia 633, 5–15.
Krause-Jensen, D., Greve, T.M., Nielsen, K., 2005. Eelgrass as a bioindicator under
the European water framework directive. Water Resources Management 19,
63–75.
Linke, P., Lutze, G.F., 1993. Microhabitat preferences of benthic foraminifera a static
concept or a dynamic adaptation to optimize food acquisition? Marine Micropa-
leontology 20, 215–234.
Martínez-Colón, M., Hallock, P., Green-Ruiz., C., 2009. Potentially toxic elements
and strategies for use of shallow water benthic foraminifers as bio-indicators: a
review. Journal of Foraminiferal Research 39, 278–299.
Muxika, I., Borja, A., Bald, J., 2007. Using historical data, expert judgement and
multivariate analysis in assessing reference conditions and benthic ecological
status, according to the European Water Framework Directive. Marine Pollution
Bulletin 55, 16–29.
Mojtahid, M., Jorissen, F., Pearson, T.H., 2008. Comparison of benthic foraminiferal
and macrofaunal responses to organic pollution in the Firth of Clyde (Scotland).
Marine Pollution Bulletin 56, 42–76.
Morvan, J., Le Cadre, V., Jorissen, F., Debenay, J.-P., 2004. Foraminifera as potential
bio-indicators of the Erika oil spill in the Bay of Bourgneuf: field and experimen-
tal studies. Aquatic Living Resources 17, 317–322.
Murray, J.W., 2000. The enigma of the continued use of total assemblages in eco-
logical studies of benthic foraminifera. Journal of Foraminiferal Research 30,
244–245.
Murray, J.W., 2006. Ecology and Applications of Benthic Foraminifera. Cambridge
University Press, Cambridge, 426 pp.
Murray, J.W., Bowser, S.S., 2000. Mortality, protoplasm decay rate, and reliability
of staining techniques to recognize ‘living’ foraminifera: a review. Journal of
Foraminiferal Research 30, 66–70.
Nielsen, K., Sømod, B., Ellegaard, C., Krause-Jensen, D., 2003. Assessing reference
conditions according to the European Water Framework Directive using mod-
elling and analysis of historical data: an example from Randers Fjord, Denmark.
Ambio 32, 287–294.
Niemistö, L., 1974. A gravity corer for studies of soft sediments. Havforskningsinst
Skr. Helsinki 238, 33–38.
Nigam, R., Saraswat, R., Panchang, R., 2006. Application of foraminifers in eco-
toxicology: retrospect, perspect and prospect. Environmental International 32,
273–283.
Oksanen, J., Blanchet, F.G., Kindt, R., Legendre, P., O’Hara, R.G., Simpson, G.L., Solymos,
P., Stevens, M.H.H., Wagner, H., 2010. Vegan: Community Ecology Package.
Peres-Neto, P.R., Jackson, D.A., 2001. How well do multivariate data sets match? The
advantages of a Procrustean superimposition approach over the Mantel test.
Oecologia 129, 169–178.
Pinto, R., Patricio, J., Baeta, A., Fath, B.D., Neto, J.M., Marques, J.C., 2009. Review
and evaluation of estuarine biotic indices to assess benthic condition. Ecological
Indicators 9, 1–25.
Rosenberg, R., Blomqvist, M., Nilsson, H.C., Cederwall, H., Dimming, A., 2004. Marine
quality assessment by use of benthic species-abundance distributions: a pro-
posed new protocol within the European Union Water Framework Directive.
Marine Pollution Bulletin 49, 728–739.
Schönfeld, J., Spezzaferri, S., Jorissen, F., Alve, E., Korsun, S., 2011. The use of
foraminifera for bio-monitoring of marine environments—results and rec-
ommendations from the FOBIMO workshop. In: Bak, M., Kaminski, M.A.,
Waskowska, A. (Eds.), Integrating Microfossil Records from the Oceans and
the Epicontinental Seas. The Grzybowski Foundation, Grzybowski Foundation
Special Publication No. 17, p. 125.
Schröder, C.J., Scott, D.B., Medioli, F.S., 1987. Can smaller benthic foraminifera be
ignored in paleoenvironmental analyses? Journal of Foraminiferal Research 17,
101–105.
Scott, D.B., Medioli, F.S., Schafer, C.T., 2001. Monitoring in Coastal Environments
Using Foraminifera and Thecamoebians Indicators. Cambridge University Press.
Shannon, C.E., Weaver, W., 1963. The Mathematical Theory of Communication. Uni-
versity of Illinois Press.